Data and AI: The Ethics of Intelligence

AI computer scientist using notebook for business data analysis

Artificial intelligence (AI) has proven its value to companies. It can reduce errors, improve processes, transform productivity, speed up decision-making, and automate mundane processes across all industries.  

 

When it comes to specialization, it offers even deeper value, providing companies with customisable capabilities that can refine manufacturing, improve safety in mining, redefine collaboration in business, and so the use cases continue. However, it’s abilities (so-called “intelligence”) are tempered by ethical issues and risks. 

 

It is my view that before a company can truly apply AI, it would need to ensure that there is the knowledge, in other words, the prerequisite collection of skill and understanding to do so.  

 

I agree with the view expressed by Jonathan Eyres in his article “The Four Quadrants of Knowledge” that knowledge exists in four quadrants.  

 

In this blog I will endeavour to show that an understanding of these quadrants of knowledge have value when it comes to the use of AI in our industries. 

 

This brings me to the first quadrant. 

 

Quadrant 1: “Knowing What We Know” 

 

“Knowledge is power” 

 

The more knowledge there is about the benefits (but also the risks) of how emerging technologies such as AI can supplement the people, processes, and existing technologies of your company, the better! 

 

Nobody understands this more than the hardworking compliance and support teams that are constantly working on updating and mitigating internal and external vulnerabilities. 

 

It is safe to say that this journey and education process is never truly complete.  

 

Quadrant 2: “What we don’t know we know” 

 

This might seem to be a confusing principle or worse, a terrible typo. But it is not! 

 

This refers to an instance where a company or its stakeholders are only becoming actively aware of something, once an event occurs that compels the company to investigate it.  

 

An example would be a breach in security: Such as a malicious phone call or email sent to a company stakeholder. 

AI’s capabilities are as effective in creating new forms of attack as they are in protecting against them. Deep fakes that leverage AI to perfectly mimic human voices and faces are becoming increasingly challenging to detect and prevent. 

In such an instance, the company had all the information needed to respond to the data breach, but due to either a lack of knowledge or training, the employee was unaware of the breach, therefore did not report the breach and the team did not act on it until after the fact. 

This is why training and well-enforced security protocols are so important. 

 

Quadrant 3: “Know what you do not know”

 

“Understanding Risks” 

 

Risk is inherent in every technology: From a poorly integrated CRM platform that uses the wrong data to skew customer engagement through to an AI model that has not been correctly trained and managed.  

 

In other words: There is always the risk that the solution will not benefit the company or its people. 

 

If that is not scary enough, there are also the risks or “AI pitfalls or quicksand” such as bias, accountability and transparency that require deft navigation to avoid being sucked in.  

 

It’s a sentiment shared by another report published by USC Annenberg which also highlighted the importance of transparency. It pointed out that, ‘many AI algorithms, particularly deep learning models, are considered “black boxes” because they are difficult to understand or interpret.’ This is further complicated by privacy and misuse issues that evolve around AI’s accessibility and ubiquity.  

 

As was outlined in my previous blogs (“AI Bots: Identifying and Navigating the Risks Part 1 and Part 2), when used without awareness of these risks, AI can spread misinformation, infringe copyright, perpetrate harmful stereotypes, and content, and have a negative impact on your company’s reputation.  

 

Therefore, “If knowledge is power, knowing what we don’t know is wisdom.” 

Adam Grant 

 

Which brings us to the next quadrant. 

 

Quadrant 4: “What we don’t know what we don’t know” 

 

“Choose the Right Partner” 

 

This comes down to trust. As discussed above, the risks of a lack of transparency and bias as a two-fold combination, can make AI a less than trustworthy ally when it comes to providing neutral and accurate insights to the company.  

 

In a recent article published by The California Management Review from Berkley, the authors suggested that the responsibility for AI and its ethical governance and use lies in the hands of decision-makers. Accountability is defined by those who chose the technology and its use cases.  

 

There are several conversations around how AI should be used and what ethical guidelines should shape this use. 

 

This is where a trusted partner becomes essential. Your company can reap the benefits of AI, and your company can integrate it seamlessly within its operations. But for this the right partner is essential.  

 

The right partner will have not only the specialized knowledge that your company does not have but will also prioritize ethics in AI and recognize the importance of implementing solutions that address and mitigate the risks.  

 

For example: A partner that understands that as part of the AI’s training, its models use vast quantities of data to gain their “intelligence” and the risk that some of this data may include legacy biases that will influence these models and the data it, in turn, provide to companies and users. 

 

Mint, with our deep understanding of AI and the leading AI technologies on the market, can help your company implement AI by following a strategic, methodical approach that sets your company up for ethical success. The goal is to benefit from AI within an ethical, transparent framework that delivers reliable and trusted results. Mint shares this goal. 

  

AI can have a powerful and positive impact on your company with the right strategy in place. When you know how your company can benefit from a customized, relevant AI integration and you find the right partner, then you can create a process that addresses every risk and ticks every box.